The future of tropical species in secondary forests: A quantitative review

Deforestation and degradation of tropical old-growth forests has the potential to cause catastrophic species extinctions. In this review, we assess whether regenerating secondary forests (SF) can support species typically found in old-growth forest (OG) and so prevent extinctions. We review 65 studies that compare faunal diversity in SF and corresponding OG, and compare the similarity of species composition both within and between these two forest types using the Sorensen, Morisita-Horn and Sorensen-Chao indices. Comparisons between traditional similarity indices and Sorensen-Chao, which minimizes sampling biases, indicated that limited sampling effort consistently reduced apparent similarity between SF and OG and that similarity between SF and OG is actually higher than previously appreciated. Similarity, which ranges from 0 to 1, varied from 0.49 to 0.92 between replicate OG sites and was correlated with similarity between SF and OG. This correlation might be an artefact of variation among studies in sampling effort, especially for vertebrates where small samples reduce apparent similarity across all comparisons, as well as a real effect of variation among studies in landscape heterogeneity and the presence of species with patchy distributions. Therefore, similarity between SF and OG cannot be interpreted without an understanding of background variation in OG. Similarity between different SF sites provided no evidence that disturbance specialists dominate SF. Similarity to OG increased rapidly with SF age; when SF was contiguous with OG; when SF was growing in small clearings; and after low intensity land uses including clearing only, shifting agriculture and tree plantations. This describes the most frequently observed tropical SF; isolated from roads and on hilly terrain unsuitable for mechanized agriculture. Thus, our analyses indicate that tropical SF can play an important role in biodiversity conservation particularly when OG forests are nearby. An important caveat remains, however. Abundance, geographic range and levels of habitat specialization are often related. Widespread, abundant, habitat generalists might dominate similarity analyses even when relatively rare OG specialists are present. Additional species-level analyses of habitat specialization will be needed before the conservation value of tropical SF is fully understood. (C) 2009 Elsevier Ltd. All rights reserved.